Professional Documents
Culture Documents
Core Memory MIT
savings in the number of vacuum tubes required and The author wishes to acknowledge the benefits de-
allow the computer pulse-repetition frequency to be rived from discussions with J. J. Burke, G. J. Gleghorn,
lowered appreciably. These results point to the fact that, W. A. Koehn, and N. L. Kreuder during the course of
rather than adapting the general-purpose computer to this investigation by the Jet Propulsion Laboratory.
every core in each xy plane, or digit plane, would bring forma tion is stored in a memory-buffer register. The
out the signal representing the stored digit. This part speed of the machine may be judged from the timing
of the read operation is destructive, and the word must diagram (Fig. 4). Note that the read-rewrite time, or
be rewritten. cycle time, is approximately 9 J.tsecs and that there are
no restraints on how frequently this cycle may be ap-
(-~ FOR ALL OTHER a-'s)
plied to the memory. It is capable, therefore, of a basic
ZERO I FOR )n ~
repetition rate of over 100 kc. Note also that the in-
formation can be available to the machine approxi-
mately 2 J.tsecs from the beginning of the cycle.
Block Schematic
Fig. 5 shows a block schema'tic of one bank of mem-
ory. Each half of the binary address in the Address
Register is translated to a 1-out-of-32 selection by a
crystal-diode Matrix and sets up a pair of "AND" gates
for x and a pair for y. The Read Flip-Flop forms a 1.5-
J.tsec pulse and sends it to the two selected Read Drivers
Fig. 3-Three-eo-ordinate selection. which supply the 0.45-ampere currents to two selection
planes. The output signal voltages from each digit piane
For the rewrite part of the cycle the selection tech- are amplified in the Sense Amplifiers and applied to
nique remains the same, except that the half currents on "AND" gates which are strobed at the optimum time by
the selection planes are now in the write polarity, which a short (0.1 J.tsec) pulse. Pulses representing ONE's then
would result in the writing of ONE's into all the cores go off to set the Buffer Register to the just-extracted
of the selected register; this writing is controllable for number. At the end of the read currents the rewrite part
each xy, or digit, plane by the use of a digit-plane wind- of the cycle starts in the same manner, except that the
ing on which may be applied a half-current of an effec- write currents have to be safely overlapped by the in-
tive polarity opposite to the write currents. The pres- hibit currents at those digit planes where ZERO's are
ence of this "inhibit" current in any digit during the to be written. This is accomplished by having the "on"
write operation leaves a ZERO; absence of the inhibit time of the Inhibit Flip-Flop overlap slightly that of the
current leaves a ONE. Write Flip-Flop. Short (1 J.tsec) currents may be applied
R. R. Everett of Massachusetts Institute of Technol- to all digit planes after the rewrite; they are called Post-
ogy has shown that these techniques may be extended Write Disturb (PWD) currents and are used to improve
into any number of co-ordinates but that the 2-co-ordi- the ONE-to-ZERO signal ratios under certain condi-
nate read and 3-co-ordinate write system just described tions. The PWD Flip-Flop forms this pulse and applies
is one of the most desirable for the Whirlwind type of it to all 17 of the Digit-Plane Drivers through "OR"
machine. inputs.
The Cores
MEMORY
ADDRESS The cores are made of General Ceramics material
REGIS TER
MF-1326B. The first bank contains their core size F-291
which has an outside diameter of 90 mils. A smaller core
X AND Y
was used in the second bank; this is F -394, 80 mils in
SELECTION
PLANES outside diameter. Single-turn switching currents are ap-
proximately 950 and 850 ma respectively, and single-
turn output voltages (at optimum strobe time) are about
0.1 volt. Switching time, under these conditions, is ap-
DIGIT
PLANES
proximately 1.2 J.tsecs.
One of the largest problems in the building of a mem-
ory of this type is the procurement of large numbers of
Fig. 4-Memory cycle. uniform cores. Core selection was made on the basis of
a series of pulse tests, approximately four per core, and
DESCRIPTION OF MEMORY
resulted in a yield for the first bank of approximately
The capaci~y of each memory bank is 1024 registers, 30 per cent of those shipped to us by the producer.
with 16 digits (plus 1 parity digit) per register. The basic (Yields have been improving materially since this first
operating mode, or cycle, consists of setting the mem- run.) The selection criterion was fundamentally that of
ory-address register to the new address and applying an upper and lower limit on the voltage output from
the read-current pulses, followed by the write currents each core when the core was excited by a sequence of
for rewriting the information just removed. The in- current pulses devised to resemble computer operation.
TO
BUFFER REG.
(17 SET LINES)
FROM
STROBE
I
1/)1
UJI
X ZI
-I 32 32 32
...J
1 FROM
0::
I- ~I BUFFER REG.
:::>1
<
::a Il.'
~:
(17 ZERO LINES)
:::>,
x 01
1
NI
...,1 32
5
INPUT
PAIRS
5 IN PUT
PAIRS
ADDR E SS REG.
Fig. 6 shows typical output-voltage pulse shapes, the 6080 tubes in the x-read group are connected together,
nominal limits within which cores were considered ac- then through a large resistor to a negative-voltage sup-
ceptable, and the strobe time at which these amplitudes ply. The cathodes of the three other groups of 6080's
were taken. The horizontal limit lines are at 90 and 120 (x-write, y-read, y-write) are all connected in a similar
mv, total pulse length is about 1.2 fJ.secs, and the vertical manner. Each group of cathodes is normally held at a
line showing the strobe time is about 0.5 fJ.seclfrom the relatively high potential by a power amplifier and is al-
start of the pulse. lowed to drop at the proper time. Thus, each 6080 acts
not only as a cathode follower but as the logical "AND"
gate shown separately in the block schematic. The large
amount of degeneration caused by the high common-
cathode resistor compensates for nonuniformity and ag-
ing changes in the characteristics of the tubes. As a re-
sult, selection-plane currents remain within very close
limits (plus or minus 4 or 5 per cent).
Fig. 6-Test-core outputs. The digit-plane driver consists of a 6080 dual triode
driven from two amplifier stages and incorporating suffi-
Basic-Circuit Types cient negative feedback from the output to the input to
The read and write currents for the selection planes of keep the current amplitude within plus or minus 5 per
the memory are supplied directly from vacuum-tube cent over expected tube, component, and power-supply
plates. A single type-6080 vacuum tube, with its sec- variations.
tions paralleled, is used to drive a given selection plane The output signal from the sensing, or read-out, wind-
in the read direction. Another such tube drives the same ing is linearly amplified from the 100-mv level up to ap-
plane in the write direction. The control grids of the proximately a 30-volt level in a single-sided, ac coupled,
6080's are driven through 6BL7 amplifiers from the wide-band feedback amplifier. The signal is then recti-
crystal-matrix output lines. The cathodes of all of the fied and applied to the suppressor grid of a 7AK7 gate
tion pattern and repetition rate were controllable to instant the Read Flip-Flop is pulsed by the Start Read
some degree by the program being run; a program which pulse. The three curves are for three values of selection-
seemed to give the most adverse pattern and rate was plane driving current, two extremes and one near op-
designed and used during most of the testing. timum. A wide operating region is again indicated.
The tests were made on the Memory Test Computer,
a high-speed, 16-digit, parallel machine of the Whirl- STROBE TIMING, I'sec
13 14 15 16 17 IB 19 20 21 22
wind type. The machine has a parity-checking system o
which computes whether each 16-digit word to be stored
contains an odd or an even number of ONE's, stores the -10
used.
-60
The bias bounds of the sense gates' suppressor grids
were chosen as a very convenient measure of the quality Fig. 13-Bias bounds versus strobe time.
of the memory output. The upper bound (least bias) is
the point at which errors occur because the gate is mis- Computer OPeration
taking the largest ZERO output for a ONE, at the lower The first bank of memory has been in use in Whirl-
bound (most bias) errors occur because the gate mis- wind since mid-August, the second since September 5.
takes the smallest ONE output for a ZERO. The bias There has been a steady improvement in their operation
difference, in volts, is a direct measure of the voltage as the installation, which was an extremely hurried one,
difference at str<.>be time between the smallest ONE and has been gradually cleaned up and made permanent
the largest ZERO. and, also, as the process of debugging these relatively
new equipments proceeded. The two banks have not
DRIVING CURRENTS,ma
(x,y, READ AND WRITE) quite been brought to an equal degree of reliability; this
0 100 200 300 400 500 600 may be due, in part, to the fact that the cores in the first
bank were not selected as carefully as those in the sec-
~
SENSE Z=400ma
GATE -10 I z = 450 rna ond so that output ONE/ZERO ratios are not as large.
The demands on the Whirlwind computer are heavy,
~~')
BIAS.
-20
VOLTS and only a few hours a month are available for further
(SUPP
GRID,
-30 PROGRAM: MP27("INCHWORM")
development work on its memory.
7AK7)
AMBIENT TEMPERATURE: 88 0 F ,jJ"
Parity alarms occurred, at first, about 3 or 4 times
-40
DATE JUNE 24,1953
per week; at this writing (November 27) there has not
-50 been a parity alarm for four weeks. This comes to about
-60L-------------------------------------~
460 hours of useful operation or, assuming a 30-p,sec
average order time and 2 accesses per average order, it
Fig. 12-Bias bounds versus drive currents.
comes to slightly over 100 billion word accesses with
Fig. 12 shows the bias bounds for all 17 sense gates each access parity checked and no error detected.
as a function of the selection-plane driving-current am- The exact nature of the errors which do occur is, as
plitudes (x, y, read, and write). The program used was yet, not known. It is hoped that further work on the
the so-called "inchworm" in which 16 words of instruc- system will shed more light on the problem as well as
tions "bootstrap" themselves around the 1024 registers reduce the error rate yet further.
of the memory. The ambient temperature was recorded
CONCLUSION
at approximately 88 degrees Fahrenheit, about 15 de-
grees higher than what is now believed to be optimum. The test results and experience obtained thus far on
Two curves are shown, one for digit-plane currents set at the two 32 by 32 by 17 banks of magnetic-core memory
400 ma and the other at 450 mao The enclosed areas indi- now operating as the internal memory of the Whirlwind
cate how much the safe operating point of the memory machine indicate high promise for this type of storage.
bank may wander; recent circuit and adjustment im- Reasonable engineering extrapolations of the results are
provements have enlarged these enclosed areas some- being used in present work on a 64 by 64 by 17 bank
what. Fig. 13 shows the bias bounds as a function of the which is expected to be in operation in the Memory Test
timing of the strobe pulse. Time is measured from the Computer by January 1954.